257

Jourraal of Chromatography, 421 (1988) 257-267

Biomedical Applications
Elsevier Science Publishers B.V., Amsterdam - Printed in The Netherlands

CHROMBIO. 4132

APPLICATION OF HIGH-PERFORMANCE LIQUID

CHROMATOGRAPHY OF SOME ANTIBIOTICS IN CLINICAL
MICROBIOLOGY

J. KNOLLER, W. KGNIG*, W. SCHONFELD, K.D. BREMM and M. KOLLER

Lehrstuhl fiir Medizinische Mikrobiologie und Immunologie, AG Infektabwehrmechanismen, Ruhr

Universitiit, Universitiitsstrasse 150,463O Bochum (F.R.G.)

(First received July 7th, 1987; revised manuscript received January 25th, 1988)

SUMMARY

During recent years high-performance liquid chromatography has become an excellent tool for the
determination of antibiotics in biological fluids. Compared with biological assays, the major benefits
of this method are specificity and rapidity. In particular, the determination of biologically inactive
metabolites emphasizes that this technique plays an outstanding role for the analysis of antibiotics.
This paper describes how the method can be used in the analysis of several antibiotics and demon-
strates the efficacy of this method for clinical microbiology. Methods for the determination in bio-
logical fluids of acylaminopenicillins (axlocillin, mezlocillin, piperacillin and aspoxicillin), quinolones
(ciprofloxacine, norfloxacine and ofloxacine), a penem (imipenem) and a cephalosporin (cefixime)
are summarized. Furthermore, their application to in vitro studies and their trial in clinical studies
are described.

INTRODUCTION

In the past ten years new techniques based on analytical methodology have
been established in microbiology [ 11. High-performance liquid chromatography
(HPLC) in particular has become an important tool for routine determination
of antimicrobial agents in body fluids. The major benefits of HPLC are specific-
ity, rapidity and sensitivity. A further advantage from the pharmacokinetic point
of view is its potential for the detection and quantitation of metabolic degradation
products, such as penicilloic and penilloic acid.
Owing to the short time required for HPLC analysis and its great potential for
the separation and detection of antibiotic drugs, many reports have focused on
the techniques for the determination of various antimicrobial drugs in biological
fluids and tissues [l-5].
This paper summarizes the results of our reversed-phase HPLC studies of an-

0378-4347/88/$03.50 0 1988 Elsevier Science Publishers B.V.

258

timicrobial agents and describes the development of standard methods as well as

their application and validation by clinical studies. Degradation studies are dealt
with only briefly because they have been reported elsewhere [ 5-131.

EXPERIMENTAL

Reagents
Azlocillin and mezlocillin, together with the appropriate penicilloates and pen-
illoates, and ciprofloxacine were kindly provided by Bayer (Leverkusen, F.R.G. ).
Norfloxacine, imipenem and cilastatin were kindly provided by Merck Sharp &
Dohme (Munich, F.R.G. ) and aspoxicillin and cefixime by E. Merck (Darmstadt,
F.R.G.). Ofloxacine was purchased from Hoechst (Frankfurt, F.R.G.) andpiper-
acillin from Cyanamide-Lederle (Munich, F.R.G.).
Methanol, dipotassium hydrogenphosphate and phosphoric acid were pur-
chased from Riedel de Haen (Seelze, F.R.G. ) and acetonitrile was purchased from
Baker Chemicals (Gross-Gerau, F.R.G.).

Buffer solutions
Soerensen buffer consisted of 66.6 mM dipotassium hydrogenphosphate ad-
justed to pH 7.40 with 66.6 mMpotassium dihydrogenphosphate.
The phosphate-buffered saline (PBS) comprised 120 mM sodium chloride-10
mM disodium hydrogenphosphate dihydrate-3 n&f potassium dihydrogenphos-
phate adjusted to pH 7.40.
Phosphate buffer 1, for /I-lactam antibiotics, was 57.4 mA4 dipotassium hydro-
genphosphate adjusted to the appropriate pH with phosphoric acid.
Phosphate buffer 2, for oxoquinolinecarboxylic acids, was 15 m&f phosphoric
acid adjusted to pH 3.0 with tetrabutylammonium hydroxide. This buffer can be
replaced by 4.4 mM tetrabutylammonium hydrogensulphate with the same pro-
portions of methanol and acetonitrile as indicated for the analysis of the oxo-
quinolinecarboxylic acid derivatives.
Phosphate buffer 3, for imipenem, was 15 m&f phosphoric acid adjusted to pH
7.0 with tetrabutylammonium hydroxide.

Study design
Degradation studies in buffer solutions were performed in Soerensen buffer.
Studies of the pH dependency were performed in borate buffer after Theorell and
Stenhagen [ 22 1.

Sample preparation
Buffer, serum and plasma samples (1:2up to 1:lO)as well as urine samples
(1:lO up to 1:lOO) were diluted with Soerensen buffer (pH 7.40) or with 0.03 A4
phosphoric acid (in the case of the oxoquinolinecarboxylic acid derivatives). After
centrifugation, 20 ~1 of the supernatant were subjected to HPLC.
The human lung and gut samples were cut with a scalpel, placed in a glass vial
with l-3ml of Soerensen buffer and homogenized in an Ultra-Turrax (Janke &
Kunkel, Staufen, F.R.G.) for lo-20 s. After centrifugation (three times at 9600
g for 5 min) ,20 ~1 of the resultant supernatant were subjected to HPLC.
The human chondral tissue samples were also cut with a scalpel into very small
pieces, placed in a glass vial containing 3-6 ml of Soerensen buffer and homoge-
nized in an Ultra-Turrax (Janke & Kunkel) in an ice-bath for 2-3 min. After
centrifugation (four or five times at 9600 g for 5 min), the supernatant (100 ~1)
was injected onto the HPLC column.
Human pleural samples were diluted with Soerensen buffer and homogenized
with a Sonifier W185 (Branson Ultrasonics, Plainview, NJ, U.S.A.). After cen-
trifugation, 20 ,ul of the supernatant were injected.

Apparatus
HPLC was performed with a Constametric III-G pump, a variable-wavelength
UV detector (Spectromonitor D ), an LDC 301 computing integrator (LDC-Mil-
ton Roy, Hasselroth, F.R.G. ) and an SP 8780 XR autosampler (Spectra-Physics,
Darmstadt, F.R.G.). The column (200 mmx4 mm I.D.) was packed with the
reversed-phase material Nucleosil C,, (5 pm) (Macherey and Nagel, Dtien,
F.R.G.).

Detection modes
Table I summarizes the detection procedures for the antibiotic drugs analysed
in our laboratory.

HPLC procedure
The solvent system consisted of a mixture (Table II) of methanol or acetoni-
trile and phosphate buffer (adjusted to the appropriate pH with phosphoric acid)
as listed under Buffer solutions. After mixing, the solvent was again adjusted with
phosphoric acid to the final pH value. The mobile phases were degassed and de-
livered at a flow-rate of 1 ml/min at room temperature. After sample preparation,
20 ~1 of the centrifuged supernatants were applied to the column. Column per-
formance was frequently checked by analysing standards after every twentieth
serum or urine sample. About 250-300 samples could be analysed with one column.
Standards (5,25,50,100 and 250 pg/ml) of the penicillins (and their metab-
olites) , imipenem and cefixime were prepared by serial dilution of a stock solution
(prepared daily) containing 1 mg/ml drug with Soerensen buffer (pH 7.40).
Quinolone standards (250, 100, 50, 25, 10, 5 and 2.5 ng/ml) were prepared by
serial dilution of a stock solution (1 mg/ml) with 0.03 M phosphoric acid. Cali-
bration curves were obtained by plotting the peak area against the concentrations
of the standard solutions. The linear relationship between the peak area and the
antibiotic concentration ranged between 1 and 250 pg/ml (10 and 250 ng/ml for
the quinolones) with a correlation coefficient of 0.998. The detection limits were
as follows: 0.1 pug/ml for the acylaminopenicillins together with their penicilloic
acid, 0.8 pug/ml for the appropriate penilloic acids, 0.5 pg/ml for aspoxicillin, 0.1
pg/ml for cefixime, 0.3 pg/ml for imipenem, 2.5 ng/ml for ciprofloxacine and
norfloxacine and 10 ng/ml for ofloxacine. All samples were stored at - 70°C until
analysis and kept cold at 4°C prior to injection.
260

TABLE I

DETECTION MODES USED

Antibiotic Detection Wavelength (nm 1

Azlocillin uv 220
Mezlocillin uv 220
Piperacillin uv 220
Aspoxicillin uv 220
Cefixime W 230
Imipenem W 313
Ciprofloxacine Fluorometry ex., 278, em., 446
Norfloxacine Fluorometry ex., 278, em., 446
Ofloxacine Fluorometry ex., 278, em., 446

TABLE II

MOBILE PHASES USED

After mixing, the solvent was adjusted with phosphoric acid to the appropriate pH value. For/%lactam
antibiotics, phosphate buffer 1 was used, and for oxoquinolinecarboxylic acids, phosphate buffer 2
was used.

Antibiotic Eluent PH
Azlocillin 16% Acetonitrile 5.0
Mezlocillin 24% Acetonitrile 4.0
Piperacillin 23% Acetonitrile 5.2
Aspoxicillin 8% Methanol 5.8
Cefixime 15% Methanol 5.2
Imipenem Phosphate buffer 3 7.0
Ofloxacine 13% Methanol and 7% acetonitrile 3.0
Ciprofloxacine 13% Methanol and 7% acetonitrile 3.0
Norfloxacine 13% Methanol and 7% acetonitrile 3.0

Peaks were identified by their retention times. The concentrations were deter-
mined by external standardization. Data obtained for mezlocillin should give in-
formation about precision and accuracy of the developed HPLC methods and
should serve as an example for the other penicillins. The linear relationship be-
tween the peak height (as well as peak area) and the mezlocillin concentration
ranged between 2.5 and 250 ,ug/ml (correlation coefficient 0.999) with the follow-
ing coefficients of variation (C.V. ): 2.30% at a cefixime concentration of 2.5pg/ml;
1.40% at 100 lug/ml; 0.8% at 250 ,ug/ml.

RESULTS AND DISCUSSION

Acylaminopenicillins
The following degradation studies of the acylureidopenicillins azlocillin, mez-
locillin and piperacillin were performed.
Mezlocillin (which was analysed as representative of the other penicillins ) was
 261

DEGRADATION OF ACYlJJKlOOPENlClLLlH IN SERUM AT 37O C

1
AX0
MEZLO c

‘IP

\ .# -.,
1_- 2s 0.c _- --
0 IO ZOmin 0 10 2Omin 20min -ii- 10 20min 0 5 10 15min 0 5 10 15min

OH 24 H

Fig. 1. Representative chromatograms for the acylureidopenicillins. (A) Serum samples spiked with
azlocillin (50 ,ug/ml), (B) mezlocillin (100 pg/ml) and (C) piperacillin (50 &ml), before and after
a 24-h incubation period in serum. Peaks: PC = penicilloate; PL =penilloate.

degraded as follows [ 71: 50% of the mezlocillin was degraded in serum at - 20’ C
after four weeks and up to 72% after six months; at 4’ C ca. 55% was decomposed
after seven days and up to 98% after four weeks. In buffer solution (pH 7.40) a
higher stability was observed:
262

Azlocillin and the metabolites penicilloate and penilloate were determined in

serum and lung tissue [ 111. At 30 min before lung resection eighteen patients
(fifteen males, three females; aged between 45 and 65 years) were treated with 5
g of azlocillin intravenously) ; all patients suffered from lung cancer. After 30 min
a mean azlocillin serum level of 322 pg/ml (n = 18) was determined, the half-life
of azlocillin was 77 min. Penicilloate was detected up to 9% (of the detected
azlocillin), and the penilloate concentration remained between 1 and 5 pg/ml.
The amount of azlocillin in the lung was 92 ,ug/g of tissue, which is a sufficient
concentration for therapy. The penicilloate amounted to 14% of the detected
azlocillin in frozen tissue samples and 4% in samples that were analysed on the
day of resection.
In colon surgery antibiotic prophylaxis is in general use. All patients (n= 11)
were treated first with 0.5 g of metronidazol (20-min infusion) followed by a 20-
min infusion with 4 g of azlocillin. Azlocillin serum levels between 219 and 343
pg/ml were determined after 30 min. In the gut samples, the azlocillin concentra-
tions were between 18.3 and 42.4 ,ug/g of tissue, indicating that sufficient pene-
tration of the antibiotic into the tissue had occurred.
Prophylactic treatment with azlocillin was carried out under surgery because
bacterial infection of chondral tissue [14] after surgical intervention is often
followed by the loss of joint function. At 45, 60 and 120 min after intravenous
infusion of 75 mg of azlocillin per kg body weight to ten patients (4-10 years old)
undergoing thoracoplastic surgery, the cartilage (0.5-3 g) and serum samples
were withdrawn and analysed by HPLC. In serum, the concentrations of azlocil-
lin were ca. 478 -t 126 pg/ml after 45 min, decreasing to 120 2 46 pug/ml after 120
min. Tissue levels of 245 16 ,ug/g (45 min) and 29 & 21 ,ug/g (120 min) were
measured.
Besides investigating whether a sufficiently high drug concentration is pro-
duced within the infectious focus, the analysis of antibiotics is undertaken to
assess whether the chemotherapy is reasonable. Therefore, information about the
concentration of mezlocillin in the pleural effusion provides an efficient treat-
ment of bronchopulmonary infections. All patients (n = 9) were treated with an
infusion of mezlocillin (10 g) for 30 min. Serum levels were determined to be
475 5 91 pg/ml (n= 7) after 30 min, falling to 310 + 87 pug/ml after 4 h. In the
pleural effusion high and long-lasting levels of mezlocillin could be detected:
68.8 + 34 pug/ml (n= 7 ) after 30 min and up to 104 t 40 pg/ml (n= 7) after 4 h.

Aspoxicillin
Aspoxicillin (TA-058) is a new semisynthetic p-lactam antibiotic with a chem-
ical similarity to other penicillins, such as mezlocillin and piperacillin. Its phar-
macokinetic behaviour, first studied in Japan, was investigated and compared
with that of piperacillin after intravenous and intramuscular injection. Plasma
and urine levels of eight healthy male volunteers (22-33 years old) were deter-
mined by HPLC (Fig. 2 ) ,
After intravenous injection of 1 g of aspoxicillin, intravenous concentrations
of 68 + 13 pg/ml were measured after 10 min, steadily decreasing with a half-life
of ca. 60 min. In comparison, piperacillin concentrations amounted to 74 -t 21
 263
A

i
0 5 10 15min

Fig. 2. Representative chromatograms of (A) buffer sample spiked with aspoxicillin (50 &ml), (B )
blank serum (diluted 1:5 with Soerensen buffer) and (C) serumsamplespikedwithaspoxicillin(50
&ml). Peak A = aspoxicillin.

pg/ml after 10 min. The recovery in urine within 24 h was ca. 97% of the dose for
aspoxicillin and 72% for piperacillin. Intramuscular administration led to plasma
aspoxicillin levels of 30 pg/ml after 30-45 min. The half-life was comparable with
that obtained after intravenous injection, and the recovery in urine was ca. 83%
within 24 h.
We also studied the stability of aspoxicillin [lo] in serum and buffer at differ-
ent temperatures over three months. Furthermore, the degradation of aspoxicil-
lin versus piperacillin was determined in serum and buffer at 37 oC. Aspoxicillin
remained stable only at - 70” C: degradation was observed at - 20” C and 4’ C.
At 37”C, 20% of aspoxicillin was degraded in serum after 24 h, whereas pipera-
cillin was completely degraded under the same conditions, as shown above.

Imipenem
Imipenem (N-formimidoyl thienamycin) was found to have the widest anti-
microbial potency of /?-lactam antibiotics, including the third-generation
cephalosporins and the new acylaminopenicillins [ 151.
Carbapenem antibiotics were extensively metabolized by a dipeptidase (de-
hydropeptidase-I) located in the brush border membrane of mammalian kidneys.
The activity of this enzyme resulted in low recoveries in urine. Coadministration
of the selective enzyme inhibitor cilastatin markedly increased the urinary recov-
ery of imipenem. We investigated imipenem degradation and its metabolism, in-
cluding extrarenal metabolism, by HPLC [ 121.
After incubation with the microsomal kidney fraction (rabbit) at 37’ C for 2 h,
40% of the drug was degraded to two metabolites (Fig. 3 ) , which showed different
UV maxima (275 and 308 nm). We suggested that these products are substances
with an open lactam ring structure. The generation of these metabolites by the
renal enzyme was completely blocked by cilastatin, a specific dipeptidase inhib-
itor [ 161.In serum, 76.6? 3.6% of the imipenem was recovered after 1 h and
52.6? 3.7% after 2 h. This recovery could be improved by 13% by preincubation
 1

J
i2

L-
3%
A
U-J..
B
li c
jy----,
D
-n-i

Fig. 3. Representative chromatograms of imipenem (0.5 mg/ml) after incubation for 3 h at 37°C:
(A) in water; (B) in the microsomal kidney fraction; (C) as B, but with a preincubation of the
microsomal kidney fraction with cilastatin (2 ,ug/ml); (D) renal microsomal fraction without imi-
penem; (E) after incubation for 1 h at room temperature in PBS buffer. Peaks: 1 =imipenem; 2 and
3 = metabolites.

of serum with EDTA. These results indicate a possible involvement of serum

dipeptidases in the systemic breakdown of the imipenem molecule.

Cefixime
Cefixime (FK-027) is a &lactamase-resistant third-generation cephalosporin
antibiotic for oral use. It exhibits a good in vitro activity against most gram-
negative pathogens and a wide range of gram-positive organisms. The cephems
comprise a wide group of compounds containing over thirty antibiotics of medical
importance. HPLC also was used to assay these substances for drug monitoring
extensively [ 171.

131 was studied in serum and buffer during storage at various tem-
peratures (4’ C, - 20 ’ C and - 70’ C ) for three months. Furthermore, its stability
in serum, buffer and urine at 37 ‘C was studied over the time period of 24 h.
Degradation was only observed at 4°C in serum. Cefixime was stable for three
months at -70°C and even at -20°C. The data obtained at 37°C over 24 h,
showed cefixime to be stable in serum (compared with the data obtained with the
acylureidopenicillins) as well as in buffer and urine, where no degradation could
be observed.

Oxoquinolinecarboxylic acid derivatives

In recent years an increasing number of condensed carboxylic acids with an-
timicrobial activity have been developed, almost all based on a 1,4-dihydroxy-li-
 265

0 5 10 15 -i5-czx 7 G&in

Fig. 4. Representative chromatograms of (A) buffer control to which cefixime (125 pg/ml) had been
added, (B ) a blank serum (diluted 1:lO with Soerensen buffer) and (C) after one month of storage
at 4°C (serum sample to which 250 &ml cefixime had been added). Peak Ce = cef’ixime.

oxo-3-quinolinecarboxylic acid skeleton. For convenience, the abbreviated name

quinolones or oxoquinolinecarboxylic acids has been used for the members of this
antibiotic group. The biological activity of the quinolones is directed against DNA
replication: they inhibit the DNA-gyrase of bacteria.
Because the metabolites are also supposed to be active in biological assays, it
was necessary to develop a method suitable for distinguishing between the parent
substances and their metabolites. This was confirmed by Joos et al. [ 181: they
showed a good correlation between a biological assay and HPLC for ciprofloxa-
tine concentrations in serum, but they obtained significant differences between
these two methods with drug concentrations in urine. The results obtained with
the bioassay were markedly higher than the HPLC values, which might be due to
microbiologically active metabolites excreted by the kidney.
Several studies dealing with the determination of norfloxacine and ciproflox-
acine in serum [ 19,201, urine, faeces, bronchial, pleural fluids and saliva showed
no interference from other serum components or compounds. Thus chromato-
graphic identification with fluorescence detection is a very specific method for
the analysis of these substances.
Because of the broad antimicrobial spectrum of ciprofloxacine [ 211, this drug
is qualified for the prophylaxis of acute chronic bronchitis. In the course of a
clinical study its clinical efficacy and its compatibility were examined. Moreover,
we were interested in analysing the drug concentrations within the focus of in-
fection. All patients (n= 17) were treated with a 500~pg ciprofloxacine tablet
after breakfast as well as after supper. In serum, drug concentrations of 945 2 758
ng/ml (n = 6) after 30 min were determined and amounted to 3358 2 826 ng/ml
after 60 min. At 4-6 h after the first tablet, 2143 5 636 ng/ml ciprofloxacine were
measured in the sputum, decreasing to 1663 2 944 ng/ml after 8-12 h after the
first tablet had been taken.
Ciprofloxacine was also studied for chemoprophylaxis under immunosuppres-
sive therapy in patients suffering from leukaemia. Patients (n = 43) were treated
 NOR YOR CIP CIP

B @ c3

-J -l
Fig. 5. Representative chromatograms of (A) blank serum, (B) serum after 24 h incubationwith1
&g/ml norfloxacine at 37°C (C) serum of a patient treated with 500-mg norfloxacine tablet, (Dl
serum after 24 h incubation with 1 pgg/ml ciprofloxacine at 37 ’ C and (E ) serum of a patient treated
with 500-mg ciprofloxacine tablet. Chromatograms B and D each show an additional peak, with a
retention time of I5 and 18 min, respectively, which are probably degradation compounds. Peaks:
NOR = norfloxacine; CIP = ciprofloxacine.

with ciprofloxacine tablets containing either 750 mg (group A) or 500 mg (group

B). The ciprofloxacine concentrations (Fig. 5) analysed in the serum ranged
between 843 and 1433 ng/ml (group A) and between 552 and 843 ng/ml (group
B) after 90 min. They reached values between 436 and 1318 ng/ml (A) and be-
tween 431 and 561 ng/ml (B) after 180 min.
In saliva, sufficiently high concentrations of ciprofloxacine were determined:
the results obtained ranged between 244 and 420 ng/ml (B) and between 395 and
637 ng/ml (A) after 90 min decreasing to values between 178and 216ng/ml (B)
as well as 287 and 470 ng/ml (A) after 180 min.

CONCLUSION

HPLC has become a valuable adjunct to biological assays for the determination
of antibiotics in body fluids and tissues. With regard to drug monitoring and
pharmacokinetic studies HPLC has outstanding advantages, especially for the
analysis of inactive (e.g. penicilloic acids) and active metabolites (such as me-
tabolites of ciprofloxacine). The HPLC technique is suitable for all groups of
antibiotic drugs [ 1 ] such as penicillins, cephalosporins, penems, aminoglycosides
and oxoquinolinecarboxylic acids, as well asp-lactamase inhibitors, etc. As pointed
out, the HPLC assay usually requires aqueous buffer systems (sometimes con-
taining ion-pair reagents) of different pH values with various proportions of or-
ganic solvents, such as methanol and acetonitrile. In contrast to other methods
[ 31, the only sample preparation required is dilution with Soerensen buffer, The
usual detection methods for antibiotic drugs are based on UV absorption [ 31. A
variable-wavelength detector is useful in the region 210-235 nm; the absorption
is more intense within this region, though antibiotics such as acylaminopenicil-
lins, cephalosporins and chloramphenicol have absorption maxima at ca 250-270,
 267

270-280 and 345-370 nm, respectively. Fluorescence detection also plays an im-
portant role [ 211, especially for the oxoquinolinecarboxylic acid derivatives, ow-
ing to the natural fluorescence of these compounds. Thus, HPLC has become a
valuable completion to biological assays in clinical microbiology and may provide
further approaches for rapid diagnosis in microbiology.

REFERENCES

1 S. Reeves and U. Ullmann (Editors), High Performance Liquid Chromatography in Medical

Microbiology, Gustav Fischer Verlag, Stuttgart, New York, 1986.
2 I. Nilson-Ehle, J. Liq. Chromatogr., 6 (1983) 251.
3 D. Fiore, F.A. Auger, G.L. Drusano and L.J. Lesco, Antimicrob. Agents Chemother., 26 (1984)
775.
4 R. Hildebrandt and U. Gundert-Remy, J. Chromatogr., 228 (1982) 409.
5 A. Weber, K.E. Opheim, K. Wong and A.L. Smith, Antimicrob. Agents Chemother., 24 (1983)
750.
6 J. Knoller, W. Schiinfeld, K.D. Bremm and W. Kijnig, Third European Congress of Clinical
Microbiology, The Hague, May ll-14,1987, Abstract No. 520.
7 J. Kniiller, K.D. Bremm, W. Schonfeld and W. Konig, Antimicrob. Agents Chemother., 29 (1986)
527.
8 W. Schonfeld, J. KnBller, K.D. Bremm and W. Kijnig, Zbl. Bakt. Hyg. A, 260 (1985) 254.
9 J. Knoller, W. Schonfeld, K.D. Bremm and W. K&rig, XIV International Congress of Microbi-
ology, Manchester, Sept. 7-13, 1986, Abstract No. P. B7-5.
10 J. Knijller, W. Schonfeld, K.D. Bremm and W. Kijnig, Zbl. Bakt. Hyg. A, 265 { 1987) 176.
11 W. Schanfeld, J. Kniiller, K.D. Bremm, R. Ritscher and W. Kiinig, FAC, 4-6 (1985) 1445.
12 M. Kijller, J. Brom, W. ScMnfeld, J. Knoller and W. Konig, Zbl. Bakt. Hyg. A, 261 (1986) 330.
13 J. Knoller, W. Schonfeld, K.D. Bremm and W. Kijnig, J Chromatogr., 389 (1987) 312.
14 W. Schijnfeld, W. Konig, J. Knijller and H. Meier, Third European Congress of Clinical Micro-
biology, The Hague, May 11-14, 1987, Abstract No. 519.
15 N.R. Jones, Am. J. Med., 78, (Suppl. 6A) (1985) 22.
16 M. Kijller, J. Brom, M. Raulf and W. Kiinig, Biochem. Biophys. Res. Commun., 131 (1985) 974.
17 L.O. White, High Performance Liquid Chromatography in Medical Microbiology, Gustav Fischer
Verlag, Stuttgart, New York, 1986, pp. 91-121.
18 B. Joos, B. Ledergerber, M. Flepp, J.-D. Bettex, R. Liithy and W. Siegenthaler, Antimicrob.
Agents Chemother., 27 (1985) 353.
19 W. Gau, H.J. Ploschke, K. Schmidt and B. Weber, J. Liq. Chromatogr., 8 (1985) 485.
20 W. Schonfeld, J. Knoller, K.D. Bremm, A. Dahlhoff, B. Weber and W. Kiinig, Zbl. Bakt. Hyg.
A, 261 (1986) 338.
21 H.L. Muytjens, J. van der Ros-van de Repe and G. van Veldhuizen, Antimicrob. Agents Chemo-
ther., 24 (1983) 302.
22 F. Theorell and W. Stenhagen, Biochem. Z., 299 (1938) 416.